While I sort of agree with your and the parent post (I have always liked Flask with simple Janga templates) I just put a task on my personal schedule to try a project with FastHTML and MonsterUI. A long time ago I found studying JH’s abstractions for deep learning very useful, even though I did all my real work in TensorFlow. It will probably be the same with FastHTML and MonsterUI: I am likely to just use it for one fun project, but I am sure to learn a few things. And, I might use it long term. A no-lose situation because I am retired and my time is my own.